mirror of https://github.com/glouw/tinn
readme update
This commit is contained in:
parent
501321b8a2
commit
9ea4e51a38
|
@ -32,6 +32,14 @@ This gives 10 outputs to the neural network. The test program will output the
|
|||
accuracy for each digit. Expect above 99% accuracy for the correct digit, and
|
||||
less that 0.1% accuracy for the other digits.
|
||||
|
||||
# Features
|
||||
|
||||
* Portable - Runs on Windows, MacOS, Linux, and embedded chips like ARM, AVR, and Microchip
|
||||
|
||||
* Sigmoidal activation.
|
||||
|
||||
* One hidden layer.
|
||||
|
||||
# Tips
|
||||
|
||||
* Tinn will never use more than the C standard library.
|
||||
|
|
Loading…
Reference in New Issue